Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7f0ba884b358>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7f0ba565bac8>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.0.1
Default GPU Device: /gpu:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    real_input = tf.placeholder(dtype=tf.float32, shape=[None, 28, 28, image_channels], name="real_input_images_ph")
    z = tf.placeholder(dtype=tf.float32, shape=[None, z_dim], name="generator_noise_input_ph")
    learning_rate = tf.placeholder(dtype=tf.float32, name="learning_rate_ph")
    return real_input, z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

def dense(x, out_units):
    return tf.layers.dense(x, out_units, activation=None, use_bias=False, 
                           kernel_initializer=tf.random_normal_initializer(mean=0.0, stddev=0.02))

def conv2d(x, output_space):
    return tf.layers.conv2d(x, output_space, kernel_size=5, strides=2, padding="same", activation=None, 
                            use_bias=False, kernel_initializer=tf.random_normal_initializer(mean=0.0, stddev=0.02))

def leaky_relu(x, alpha=0.2):
     # non-linear activation function
    return tf.maximum(alpha * x, x)

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the generator, tensor logits of the generator).

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param image: Tensor of input image(s) - shape = (None,28,28,3)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """

    with tf.variable_scope("discriminator", reuse=reuse):
        # Use LeakyReLU activation in the discriminator for all layers but the input layer.
        # convolution > Leaky RELUs, we do not use batch_norm on the first layer of the discriminator
        conv1 = conv2d(images, 64)
        conv1 = leaky_relu(conv1)
        
        # convolution > batch normalization > Leaky RELUs
        conv2 = conv2d(conv1, 128)
        conv2 = tf.layers.batch_normalization(conv2, training=True)
        conv2 = leaky_relu(conv2)
        
        conv3 = conv2d(conv2, 256)
        conv3 = tf.layers.batch_normalization(conv3, training=True)
        conv3 = leaky_relu(conv3) # non-linear activation function
        
        flatten = tf.contrib.layers.flatten(conv3)
        
        logits = dense(flatten, 1)
        output = tf.nn.sigmoid(logits)
        
    return output, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # Use ReLU activation in generator for all layers except for the output, which uses Tanh
    with tf.variable_scope('generator', reuse= not is_train):
        x = dense(z, 2*2*512)
        x = tf.reshape(x, [-1,2,2,512])
        x = tf.layers.batch_normalization(x, training=is_train)
        x = tf.nn.relu(x) 
        # shape=(?, 2, 2, 512)
        
        t_conv1 = tf.layers.conv2d_transpose(x, 256, 5, strides=2, padding="valid", activation=None, use_bias=False)
        t_conv1 = tf.layers.batch_normalization(t_conv1, training=is_train)
        t_conv1 = tf.nn.relu(t_conv1)
        # shape=(?, 7, 7, 256)
        
        t_conv2 = tf.layers.conv2d_transpose(t_conv1, 128, 5, strides=2, padding="same", activation=None, use_bias=False)
        t_conv2 = tf.layers.batch_normalization(t_conv2, training=is_train)
        t_conv2 = tf.nn.relu(t_conv2)
        # shape=(?, 14, 14, 128)
        
        # we do not apply batch normalization to the last layer of the generator
        # also, we apply tanh instead of relu to the output layer of the generator
        output = tf.layers.conv2d_transpose(t_conv2, out_channel_dim, 5, strides=2, padding="same", activation=tf.nn.tanh, use_bias=False)

        return output


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    smooth = 0.1
    generated_image = generator(input_z, out_channel_dim, is_train=True)
    
    # we build 2 discriminator nets, one for real data one for fake data (comming from the generator)
    d_model_real, d_logits_real = discriminator(input_real, reuse=False)
    
    # reuse the variables here so that a new set of variables will not be created
    d_model_fake, d_logits_fake = discriminator(generated_image, reuse=True) 
    
    # since these are all real images, wee want the labels to be all ones, with label smoothing
    d_loss_real = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, 
                                                                         labels=tf.ones_like(d_model_real) * (1 - smooth)))
    
    # Since we want the discriminator to output 1 for real images and 0 for fake images, the logits here are all zeros
    d_loss_fake = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, 
                                                                         labels=tf.zeros_like(d_model_fake)))
    
    # for the generator loss, we use the fake logits from the discriminator with labels all ones
    generator_loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, 
                                                                            labels=tf.ones_like(d_model_fake) * (1 - smooth)))
    
    # for the discriminator, the total loss is the sum of the losses for the real and fake images
    discriminator_loss = d_loss_real + d_loss_fake
    
    return discriminator_loss, generator_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    all_vars = tf.trainable_variables()
    g_vars = [var for var in all_vars if var.name.startswith("generator")]
    d_vars = [var for var in all_vars if var.name.startswith("discriminator")]
    
    # If we don't include the update ops as dependencies on the train step, the 
    # tf.layers.batch_normalization layers won't update their population statistics,
    # which will cause the model to fail at inference time
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        discriminator_train_op = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        generator_train_op = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)

    return discriminator_train_op, generator_train_op


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """

    image_width, image_height, image_channels = data_shape[1], data_shape[2], data_shape[3]
    
    # get the inputs and learning rate placeholders
    real_input, z, lr = model_inputs(image_width, image_height, image_channels, z_dim)
    
    # get the model loss operations
    d_loss, g_loss = model_loss(real_input, z, image_channels)
    
    # get the training operations
    d_train_op, g_train_op = model_opt(d_loss, g_loss, lr, beta1)

    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_counter, batch_images in enumerate(get_batches(batch_size)):
                
                # generate a random noise vector for feeding the generator
                batch_z = np.random.uniform(-1.0, 1.0, size=(batch_size, z_dim))
                
                # run network optimizers
                d_op, d_train_loss = sess.run([d_train_op, d_loss], feed_dict={real_input: batch_images, z: batch_z, lr: learning_rate})
                g_op, g_train_loss = sess.run([g_train_op, g_loss], feed_dict={real_input: batch_images, z: batch_z, lr: learning_rate})
                
                # display the generator and discriminator loss during training
                if batch_counter % 10 == 0:
                    print("Epoch {}/{}...".format(epoch_i+1, epoch_count),
                          "Discriminator Loss: {:.4f}...".format(d_train_loss),
                          "Generator Loss: {:.4f}".format(g_train_loss))    
                
                # display some of the generator output images during training
                if batch_counter % 100 == 0:    
                    show_generator_output(sess, 20, z, image_channels, data_image_mode)

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 128
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 10

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 1/10... Discriminator Loss: 2.1497... Generator Loss: 2.7094

Epoch 1/10... Discriminator Loss: 0.8496... Generator Loss: 5.1474
Epoch 1/10... Discriminator Loss: 0.6075... Generator Loss: 5.1298
Epoch 1/10... Discriminator Loss: 0.7498... Generator Loss: 7.7806
Epoch 1/10... Discriminator Loss: 0.7493... Generator Loss: 6.1721
Epoch 1/10... Discriminator Loss: 0.6882... Generator Loss: 6.2109
Epoch 1/10... Discriminator Loss: 0.5879... Generator Loss: 5.0652
Epoch 1/10... Discriminator Loss: 0.4311... Generator Loss: 4.1596
Epoch 1/10... Discriminator Loss: 0.4735... Generator Loss: 4.7671
Epoch 1/10... Discriminator Loss: 1.0803... Generator Loss: 5.6443
Epoch 1/10... Discriminator Loss: 0.8291... Generator Loss: 7.3136

Epoch 1/10... Discriminator Loss: 0.7153... Generator Loss: 5.6055
Epoch 1/10... Discriminator Loss: 0.4818... Generator Loss: 4.1222
Epoch 1/10... Discriminator Loss: 0.6210... Generator Loss: 4.6122
Epoch 1/10... Discriminator Loss: 0.4642... Generator Loss: 4.0433
Epoch 1/10... Discriminator Loss: 0.4647... Generator Loss: 3.8858
Epoch 1/10... Discriminator Loss: 0.4468... Generator Loss: 4.1570
Epoch 1/10... Discriminator Loss: 0.4751... Generator Loss: 4.4750
Epoch 1/10... Discriminator Loss: 0.4845... Generator Loss: 4.0368
Epoch 1/10... Discriminator Loss: 0.4147... Generator Loss: 4.1504
Epoch 1/10... Discriminator Loss: 0.4557... Generator Loss: 4.1000

Epoch 1/10... Discriminator Loss: 0.4392... Generator Loss: 4.6205
Epoch 1/10... Discriminator Loss: 0.4431... Generator Loss: 4.6708
Epoch 1/10... Discriminator Loss: 0.4425... Generator Loss: 4.0920
Epoch 1/10... Discriminator Loss: 0.4825... Generator Loss: 3.3709
Epoch 1/10... Discriminator Loss: 0.4575... Generator Loss: 4.0767
Epoch 1/10... Discriminator Loss: 0.4115... Generator Loss: 4.7002
Epoch 1/10... Discriminator Loss: 0.4407... Generator Loss: 4.1966
Epoch 1/10... Discriminator Loss: 0.4349... Generator Loss: 3.8835
Epoch 1/10... Discriminator Loss: 0.5010... Generator Loss: 3.7292
Epoch 1/10... Discriminator Loss: 0.4589... Generator Loss: 4.1378

Epoch 1/10... Discriminator Loss: 0.4662... Generator Loss: 5.0908
Epoch 1/10... Discriminator Loss: 0.4052... Generator Loss: 4.0417
Epoch 1/10... Discriminator Loss: 0.4648... Generator Loss: 5.1275
Epoch 1/10... Discriminator Loss: 0.4151... Generator Loss: 3.9927
Epoch 1/10... Discriminator Loss: 0.4714... Generator Loss: 4.7260
Epoch 1/10... Discriminator Loss: 0.4528... Generator Loss: 4.2758
Epoch 1/10... Discriminator Loss: 0.5648... Generator Loss: 6.5949
Epoch 1/10... Discriminator Loss: 0.6634... Generator Loss: 3.0110
Epoch 1/10... Discriminator Loss: 0.4498... Generator Loss: 5.2171
Epoch 1/10... Discriminator Loss: 0.4380... Generator Loss: 3.2032

Epoch 1/10... Discriminator Loss: 0.4963... Generator Loss: 3.5201
Epoch 1/10... Discriminator Loss: 0.4943... Generator Loss: 3.1371
Epoch 1/10... Discriminator Loss: 0.4893... Generator Loss: 3.9378
Epoch 1/10... Discriminator Loss: 2.0105... Generator Loss: 6.3319
Epoch 1/10... Discriminator Loss: 0.5865... Generator Loss: 2.8306
Epoch 1/10... Discriminator Loss: 0.5917... Generator Loss: 3.5594
Epoch 2/10... Discriminator Loss: 0.4916... Generator Loss: 3.2416

Epoch 2/10... Discriminator Loss: 0.6856... Generator Loss: 4.7802
Epoch 2/10... Discriminator Loss: 0.7285... Generator Loss: 2.5021
Epoch 2/10... Discriminator Loss: 0.5700... Generator Loss: 3.4169
Epoch 2/10... Discriminator Loss: 0.5059... Generator Loss: 4.0731
Epoch 2/10... Discriminator Loss: 0.5436... Generator Loss: 3.9594
Epoch 2/10... Discriminator Loss: 0.5576... Generator Loss: 2.3477
Epoch 2/10... Discriminator Loss: 1.5724... Generator Loss: 11.4963
Epoch 2/10... Discriminator Loss: 0.5694... Generator Loss: 3.5645
Epoch 2/10... Discriminator Loss: 0.6935... Generator Loss: 5.2662
Epoch 2/10... Discriminator Loss: 0.4426... Generator Loss: 3.9470

Epoch 2/10... Discriminator Loss: 0.5728... Generator Loss: 3.0956
Epoch 2/10... Discriminator Loss: 0.4358... Generator Loss: 3.9131
Epoch 2/10... Discriminator Loss: 0.4914... Generator Loss: 3.5216
Epoch 2/10... Discriminator Loss: 0.4418... Generator Loss: 3.5156
Epoch 2/10... Discriminator Loss: 0.4741... Generator Loss: 2.9145
Epoch 2/10... Discriminator Loss: 0.5804... Generator Loss: 4.1312
Epoch 2/10... Discriminator Loss: 0.5757... Generator Loss: 4.7706
Epoch 2/10... Discriminator Loss: 0.4818... Generator Loss: 3.9200
Epoch 2/10... Discriminator Loss: 0.5942... Generator Loss: 1.8267
Epoch 2/10... Discriminator Loss: 0.5044... Generator Loss: 3.6143

Epoch 2/10... Discriminator Loss: 0.6795... Generator Loss: 5.9561
Epoch 2/10... Discriminator Loss: 0.4847... Generator Loss: 4.0574
Epoch 2/10... Discriminator Loss: 0.4399... Generator Loss: 3.4610
Epoch 2/10... Discriminator Loss: 0.4931... Generator Loss: 3.2205
Epoch 2/10... Discriminator Loss: 0.4214... Generator Loss: 3.8292
Epoch 2/10... Discriminator Loss: 0.4262... Generator Loss: 3.9705
Epoch 2/10... Discriminator Loss: 0.4521... Generator Loss: 3.5678
Epoch 2/10... Discriminator Loss: 0.4460... Generator Loss: 3.5175
Epoch 2/10... Discriminator Loss: 0.4101... Generator Loss: 3.7052
Epoch 2/10... Discriminator Loss: 0.4472... Generator Loss: 4.1334

Epoch 2/10... Discriminator Loss: 0.5571... Generator Loss: 5.5272
Epoch 2/10... Discriminator Loss: 0.5744... Generator Loss: 7.7009
Epoch 2/10... Discriminator Loss: 0.6200... Generator Loss: 4.7247
Epoch 2/10... Discriminator Loss: 0.4165... Generator Loss: 4.1330
Epoch 2/10... Discriminator Loss: 0.4041... Generator Loss: 4.0448
Epoch 2/10... Discriminator Loss: 0.3907... Generator Loss: 3.8866
Epoch 2/10... Discriminator Loss: 0.4694... Generator Loss: 3.9427
Epoch 2/10... Discriminator Loss: 0.4221... Generator Loss: 3.2416
Epoch 2/10... Discriminator Loss: 0.5729... Generator Loss: 4.8457
Epoch 2/10... Discriminator Loss: 0.4427... Generator Loss: 3.1447

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 128
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5



"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 25

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 1/25... Discriminator Loss: 1.4220... Generator Loss: 3.9563

Epoch 1/25... Discriminator Loss: 1.6415... Generator Loss: 6.1172
Epoch 1/25... Discriminator Loss: 1.0475... Generator Loss: 6.3207
Epoch 1/25... Discriminator Loss: 0.8140... Generator Loss: 5.5218
Epoch 1/25... Discriminator Loss: 0.7159... Generator Loss: 6.2071
Epoch 1/25... Discriminator Loss: 0.5709... Generator Loss: 4.7480
Epoch 1/25... Discriminator Loss: 0.5686... Generator Loss: 4.3513
Epoch 1/25... Discriminator Loss: 0.5587... Generator Loss: 3.8861
Epoch 1/25... Discriminator Loss: 0.6548... Generator Loss: 5.7524
Epoch 1/25... Discriminator Loss: 0.5583... Generator Loss: 3.9249
Epoch 1/25... Discriminator Loss: 0.5660... Generator Loss: 3.8573

Epoch 1/25... Discriminator Loss: 0.5929... Generator Loss: 5.6878
Epoch 1/25... Discriminator Loss: 0.6119... Generator Loss: 3.9105
Epoch 1/25... Discriminator Loss: 0.4530... Generator Loss: 4.2601
Epoch 1/25... Discriminator Loss: 0.4991... Generator Loss: 3.4850
Epoch 1/25... Discriminator Loss: 1.0394... Generator Loss: 2.3786
Epoch 1/25... Discriminator Loss: 0.5751... Generator Loss: 2.9290
Epoch 1/25... Discriminator Loss: 0.6042... Generator Loss: 4.2130
Epoch 1/25... Discriminator Loss: 0.8524... Generator Loss: 4.3197
Epoch 1/25... Discriminator Loss: 0.7439... Generator Loss: 2.2627
Epoch 1/25... Discriminator Loss: 0.5998... Generator Loss: 5.1500

Epoch 1/25... Discriminator Loss: 0.6394... Generator Loss: 4.5392
Epoch 1/25... Discriminator Loss: 1.6334... Generator Loss: 9.9709
Epoch 1/25... Discriminator Loss: 0.7621... Generator Loss: 2.8122
Epoch 1/25... Discriminator Loss: 0.8753... Generator Loss: 4.9137
Epoch 1/25... Discriminator Loss: 0.7520... Generator Loss: 4.2885
Epoch 1/25... Discriminator Loss: 0.8706... Generator Loss: 2.4064
Epoch 1/25... Discriminator Loss: 0.7812... Generator Loss: 3.0276
Epoch 1/25... Discriminator Loss: 0.8258... Generator Loss: 3.7385
Epoch 1/25... Discriminator Loss: 1.4509... Generator Loss: 2.2918
Epoch 1/25... Discriminator Loss: 0.7001... Generator Loss: 2.8818

Epoch 1/25... Discriminator Loss: 0.5507... Generator Loss: 3.0363
Epoch 1/25... Discriminator Loss: 0.7103... Generator Loss: 2.8030
Epoch 1/25... Discriminator Loss: 0.6353... Generator Loss: 2.6375
Epoch 1/25... Discriminator Loss: 0.7141... Generator Loss: 3.2918
Epoch 1/25... Discriminator Loss: 0.7204... Generator Loss: 3.5605
Epoch 1/25... Discriminator Loss: 0.9317... Generator Loss: 2.1341
Epoch 1/25... Discriminator Loss: 0.6534... Generator Loss: 2.4657
Epoch 1/25... Discriminator Loss: 0.6539... Generator Loss: 2.7512
Epoch 1/25... Discriminator Loss: 0.5784... Generator Loss: 2.4623
Epoch 1/25... Discriminator Loss: 0.4694... Generator Loss: 2.6890

Epoch 1/25... Discriminator Loss: 0.4970... Generator Loss: 3.0775
Epoch 1/25... Discriminator Loss: 1.0752... Generator Loss: 5.6098
Epoch 1/25... Discriminator Loss: 1.5679... Generator Loss: 2.5318
Epoch 1/25... Discriminator Loss: 0.6144... Generator Loss: 2.5623
Epoch 1/25... Discriminator Loss: 1.2834... Generator Loss: 4.2967
Epoch 1/25... Discriminator Loss: 0.8226... Generator Loss: 4.0023
Epoch 1/25... Discriminator Loss: 0.5535... Generator Loss: 3.0635
Epoch 1/25... Discriminator Loss: 0.9748... Generator Loss: 4.1063
Epoch 1/25... Discriminator Loss: 0.4163... Generator Loss: 3.6820
Epoch 1/25... Discriminator Loss: 0.5820... Generator Loss: 3.6227

Epoch 1/25... Discriminator Loss: 0.4152... Generator Loss: 3.5352
Epoch 1/25... Discriminator Loss: 1.6795... Generator Loss: 3.0699
Epoch 1/25... Discriminator Loss: 0.4529... Generator Loss: 3.3310
Epoch 1/25... Discriminator Loss: 0.4408... Generator Loss: 3.1872
Epoch 1/25... Discriminator Loss: 0.8252... Generator Loss: 2.3441
Epoch 1/25... Discriminator Loss: 0.5260... Generator Loss: 2.3481
Epoch 1/25... Discriminator Loss: 0.4576... Generator Loss: 3.1290
Epoch 1/25... Discriminator Loss: 1.1623... Generator Loss: 3.3589
Epoch 1/25... Discriminator Loss: 0.7713... Generator Loss: 2.8966
Epoch 1/25... Discriminator Loss: 0.9402... Generator Loss: 3.2754

Epoch 1/25... Discriminator Loss: 0.4600... Generator Loss: 2.6290
Epoch 1/25... Discriminator Loss: 0.6326... Generator Loss: 2.1903
Epoch 1/25... Discriminator Loss: 0.6809... Generator Loss: 5.9085
Epoch 1/25... Discriminator Loss: 0.5313... Generator Loss: 3.1737
Epoch 1/25... Discriminator Loss: 0.4407... Generator Loss: 3.9686
Epoch 1/25... Discriminator Loss: 0.9177... Generator Loss: 2.7437
Epoch 1/25... Discriminator Loss: 0.6299... Generator Loss: 3.3325
Epoch 1/25... Discriminator Loss: 1.1043... Generator Loss: 4.2907
Epoch 1/25... Discriminator Loss: 0.3928... Generator Loss: 4.4165
Epoch 1/25... Discriminator Loss: 0.4910... Generator Loss: 3.6149

Epoch 1/25... Discriminator Loss: 1.8228... Generator Loss: 3.5083
Epoch 1/25... Discriminator Loss: 1.2216... Generator Loss: 3.7746
Epoch 1/25... Discriminator Loss: 1.0430... Generator Loss: 2.8151
Epoch 1/25... Discriminator Loss: 1.0104... Generator Loss: 2.9558
Epoch 1/25... Discriminator Loss: 1.1097... Generator Loss: 4.6452
Epoch 1/25... Discriminator Loss: 0.5653... Generator Loss: 2.4436
Epoch 1/25... Discriminator Loss: 0.3672... Generator Loss: 4.9614
Epoch 1/25... Discriminator Loss: 0.4589... Generator Loss: 2.6571
Epoch 1/25... Discriminator Loss: 0.4783... Generator Loss: 1.9977
Epoch 1/25... Discriminator Loss: 0.3860... Generator Loss: 3.6420

Epoch 1/25... Discriminator Loss: 1.5747... Generator Loss: 2.7465
Epoch 1/25... Discriminator Loss: 1.5689... Generator Loss: 4.0670
Epoch 1/25... Discriminator Loss: 0.8797... Generator Loss: 1.4385
Epoch 1/25... Discriminator Loss: 0.5064... Generator Loss: 2.9861
Epoch 1/25... Discriminator Loss: 0.4468... Generator Loss: 4.1261
Epoch 1/25... Discriminator Loss: 0.4048... Generator Loss: 4.3885
Epoch 1/25... Discriminator Loss: 1.0640... Generator Loss: 2.4575
Epoch 1/25... Discriminator Loss: 1.4783... Generator Loss: 2.8772
Epoch 1/25... Discriminator Loss: 0.7135... Generator Loss: 1.9015
Epoch 1/25... Discriminator Loss: 0.4744... Generator Loss: 3.5682

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.